For the case of electric buses charging under the constraint of feed line capacity, an improved Double Center Particle Swarm Optimization algorithm (DCPSO) which can optimize the charging is offered to get the maximum sum of electric buses operating. Firstly, a mathematical optimization model of the minimum value of difference between daily peak and valley load of feed line is established, the initial sum of electric buses operating is assumed and the improved DCPSO is applied to optimize the charging. Secondly, according to the optimization results and the capacity constraint of the feed line, the sum of electric buses on operating is modified to re-optimize the charging. The modified sum of electric buses on operating will make the load curve's peak approach the feed line capacity, which will find the maximum sum of electric buses on operating at last. The improved DCPSO makes particles increase 5 motion routes based on DCPSO, which improves search range and prevents particle from premature. In order to improve the optimization speed, the initialization of particles is determined according to the distance between the daily load curve and the feeder capacity constraint. An example of typical feed line of a southern city is applied to simulate in the end, the result indicates the scheduling is more optimal.